To what degree is the current code too slow? Is it a factor of a million? a thousand? two? 10 percent?

How massive is the data you are proceessing? benchmarking the current code on a thousand lines, ten thousand lines, a million lines of your data. What is the projected time for processing the entire input?

Does there already exist a progrgam to process the files? Are there better algorithms? Is there a way to avoid proceessing the file at all?